Anand Raj · Data Analysis
Tamil Nadu Assembly Elections · 2026 · Investigative
TN 2026 · Dummy Candidate Investigation · Post-Results
The question everyone was asking after the candidate lists dropped

329 Dummies.
How Many Actually Mattered?

The honest answer depends on how strict your definition is.

Three weeks before polling, 329 suspect candidate pairs were flagged: major-alliance nominees who shared a name with another candidate in the same constituency. The thesis: at least some were deliberate vote-splitters. Today the votes are in. So did it work?

Strict test (EXACT + NEAR_FULL matches only)
0
Seats where dummy votes provably covered the margin. In the cleanest test, zero seats turned on engineered name confusion.
Loose test (all flagged dummies including WORD_MATCH)
2
Razor-thin TVK losses where dummy vote totals exceeded the winning margin. Both under 0.35% of votes polled.
329
Flagged Pairs
263
Actually Contested
152
Majors Who Lost
~200
Median Dummy Votes
Dummy Classification · 3 tiers
Exact Match
64 pairs
Identical names after normalising prefixes and initials. Most likely to be engineered. Median 174 votes polled.
Near-Full Match
36 pairs
85%+ string similarity. Still high confidence. Median 210 votes polled. 0 strict-consequential.
Word Match
163 pairs
One shared significant name fragment. Noisiest signal. The 2 near-misses are both in this tier. Median 197 votes.
A
Anand Raj
Data journalist · ndranandraj.com
May 2026 Results Day Analysis
Tamil Nadu 2026 Elections Dummy Candidates Electoral Strategy Name Confusion Data Investigation ECI Results